M. J. JESSUP 1, R. D. LAW 1, M. P. SEARLE 2 & M. S. HUBBARD 3. (

Transcription

1 Structural evolution and vorticity of flow during extrusion and exhumation of the Greater Himalayan Slab, Mount Everest Massif, Tibet/Nepal: implications for orogen-scale flow partitioning M. J. JESSUP 1,. D. LAW 1, M. P. SEALE 2 & M. S. HUBBAD 3 1 Department of Geosciences, Virginia Tech, Blacksburg, Virginia 24061, USA ( mjessup@vt.edu) 2 Department of Earth Sciences, Oxford University, Oxford, OX1 3P, UK 3 Department of Geology, Kansas State University, Manhattan, Kansas 66506, USA Abstract: The Greater Himalayan Slab (GHS) is composed of a north-dipping anatectic core, bounded above by the South Tibetan detachment system (STDS) and below by the Main Central thrust zone (MCTZ). Assuming simultaneous movement on the MCTZ and STDS, the GHS can be modelled as a southward-extruding wedge or channel. New insights into extrusionrelated flow within the GHS emerge from detailed kinematic and vorticity analyses in the Everest region. At the highest structural levels, mean kinematic vorticity number (Wm) estimates of (c % pure shear) were obtained from sheared Tethyan limestone and marble from the Yellow Band on Mount Everest. Underlying amphibolite-facies schists and gneisses, exposed in ongbuk valley, yield Wm estimates of (c % pure shear) and associated microstructures indicate that flow occurred at close to peak metamorphic conditions. Vorticity analysis becomes progressively more problematic as deformation temperatures increase towards the anatectic core. Within the MCTZ, rigid elongate garnet grains yield Wm estimates of (c % pure shear). We attribute flow partitioning in the GHS to spatial and temporal variations that resulted in the juxtaposition of amphibolite-facies rocks, which record early stages of extrusion, with greenschist to unmetamorphosed samples that record later stages of exhumation. The.2500 km length of the Himalayan orogen is cored by a suite of north-dipping metamorphic rocks (the Greater Himalayan Slab; GHS), that are bounded above and below by the normal-sense South Tibetan detachment system (STDS) and reverse-sense Main Central thrust zone (MCTZ), respectively (Figs 1 & 2). Assuming simultaneous movement along these crustal-scale bounding shear zones (see review by Godin et al. 2006b), the GHS is often modelled as a north-dipping wedge or channel of mid-crustal rocks that was extruded southward from beneath the Tibetan plateau (Fig. 2) beginning in early Miocene time (e.g. Burchfiel & oyden 1985; Burchfiel et al. 1992; Hodges et al. 1992). Although consensus on this general concept of extrusion during crustal convergence exists, and a range of orogen-scale kinematic and thermal mechanical extrusion models have been proposed, surprisingly little research has focused on quantifying the kinematics (vorticity) of flow within the slab and its potential causal relationship with progressive exhumation of the GHS. The use of vorticity analysis to quantify flow within sheared rocks has proven to be a useful tool for quantifying the nature and distribution of flow regimes within a range of tectonic settings including contractional (e.g. Simpson & De Paor 1997; Xypolias & Doutsos 2000; Xypolias & Koukouvelas 2001, Xypolias & Kokkalas 2006), extensional (Wells 2001; Bailey & Eyster 2003) and transpressional (Wallis 1995; Klepeis et al. 1999; Holcombe & Little 2001; Bailey et al. 2004) regimes. Vorticity analysis enables estimation of the relative contributions of pure and simple shear, yielding important constraints for GHS extrusion models. Identification of a pure shear component is critically important because such flow would result in: (1) thinning and transport-parallel extension of the slab itself, and (2) an increase in both strain rates and extrusion rates relative to strict simple shear. Attempts to quantify From: LAW,. D., SEALE, M. P.& GODIN, L. (eds) Channel Flow, Ductile Extrusion and Exhumation in Continental Collision Zones. Geological Society, London, Special Publications, 268, /06/$15.00 # The Geological Society of London 2006.

2 380 M. J. JESSUP ET AL. 70 E 80 E 90 E 100 E 110 E 40 N Hindu Kush Pamir Karakorum Kohistan Zanskar MMT Ladakh Karakorum fault 78 E 30 N 20 N Indian Plate Tibetan Plateau Gongsampa transect MFT MBT ZSZ Shivling 10 N a Sutlej Valley MCT STDS Indus Suture Zone Everest Transect 75 E km Transhimalayan batholith Kohistan arc Indus Tsangpo suture zone Tibetan Zone Greater Himalayan Slab Lesser Himalaya MBT MCT MBT STDS MMT Main Mantle thrust Main Zanskar backthrust Khatmandu Darjeeling STDS South Tibetan Detachment system MCTZ Main Central thrust zone 90 E MBT Main Boundary thrust STDS MCT MBT Bhutan Fig. 1. Simplified geological map of the Himalaya including the distribution of the main lithotectonic elements of the orogen. The location of the Everest transect and significant geographical locations and areas referred to in the text are indicated. Inset (a) is a digital elevation model displaying the range of elevation for SE Asia (dark grey ¼ low; light grey ¼ high; white ¼ ocean). The Himalaya marks the transition from high elevations of the Tibetan plateau to the lowlands of the Indian plate. MBT, Main Boundary thrust; MCTZ, Main Central thrust zone; MFT, Main Frontal thrust; MMT, Main Mantle thrust; STDS, South Tibetan detachment system; ZSZ, Zanskar shear zone. flow within the GHS that accommodated this southward extrusion are limited to: (1) a single transect through the lowermost 900 m of the GHS exposed in the Sutlej valley of NW India (Fig. 1; Grasemann et al. 1999; Vannay & Grasemann 2001); (2) preliminary results from the top of the GHS exposed in the ongbuk valley on the north side of the Everest massif, Tibet (Law et al. 2004); and (3) preliminary results from the middle of the GHS in the Bhutan Himalaya (Carosi et al. 2006). Quantifying and characterizing flow within the GHS is important for development of more realistic models for evolution of the Himalaya, particularly those that propose a synergistic interplay between extrusion, erosion and exhumation (e.g. Beaumont et al. 2001, 2004, 2006; Hodges et al. 2001; Grujic et al. 2002; Jamieson et al. 2004, 2006; Hodges 2006). The topographic relief of the Everest massif, Tibet/Nepal (Fig. 1), provides a window into mid-crustal processes responsible for extrusion of the GHS, and a particularly appropriate field area to test the various components of extrusion models. In this paper, we combine field-based structural analysis with detailed vorticity analyses of samples from a north south transect through the GHS in the Everest region using the rigid grain technique of Wallis et al. (1993) and Wallis (1995). Samples collected for vorticity analyses are from a variety of structural and lithologic settings, including high-altitude and summit samples collected during two pioneering climbing expeditions (1933 and 1953) on the north and south sides of Mount Everest, respectively. igid grain analysis using elongate garnet porphyroclasts quantify flow along the MCTZ and, by integration

3 FLOW PATITIONING IN THE HIMALAYA 381 N Tibet with microstructural analysis, constrain the timing of mylonite formation in relation to peak metamorphism (Hubbard 1988, 1989). Field-based structural analysis characterizes deformation in the core of the GHS where deformation temperatures exceed the upper limit for robust rigid grain vorticity analysis. As the first attempt to quantify flow in a transect across the entire GHS, many new insights into vorticity of flow emerge, including an understanding of how flow was partitioned during extrusion and exhumation of the GHS. Tectonic setting STDS STDS GHS GHS MCTZ (a) Model 1. Adapted from Burchfiel & oyden (1985) Tibet Indian Shield MCTZ Indian Shield (b) Model 2. Adapted from Grasemann et al. (1999) Tibet gneiss dome partial melt zone- STDS GHS Tibet STDS Indian lower crust GHS MCTZ Indian Shield (c) Model 3. Adapted from Grujic et al. (1996) LHS MCTZ MBT The Himalaya Tibet orogenic system has accommodated crustal convergence since initiation of collision between India and Asia at c Ma MFT Indian Shield (d) Model 4. Adapted from Beaumont et al. (2004) 40 km (V:H ~2:1) Fig. 2. (a d) Extrusion and channel flow models proposed for the evolution of the Greater Himalayan Slab (GHS). See text for detailed review of each model. GHS, Greater Himalayan Slab; LHS, Lesser Himalayan Sequence; MBT, Main Boundary thrust; MCTZ, Main Central thrust zone; MFT, Main Frontal thrust; STDS, South Tibetan detachment system. S (Searle et al. 1987; Hodges 2000; Yin & Harrison 2000; Figs 1 & 2). The Tibetan plateau encompasses an area of km 2 of subdued topography (Fig. 1, inset a), with an average elevation of c m (Fielding et al. 1994). To the south stretch the flat, low elevations characteristic of the undeformed internal margin of the Indian plate. Between lies the crest of the Himalaya, which extends for c km along-strike, contains the highest elevations in the world (8850 m), and provides exposure of mid-crustal rocks belonging to the GHS (Figs 1 & 2). The GHS forms a 5 30 km thick section of metasedimentary rocks that are intruded by leucogranite dykes and migmatized to varying degrees (Hodges 2000). The age of leucogranite crystallization (c Ma) suggests they were part of a protracted event that marks the culmination of peak metamorphism (Searle 1996; Hodges 2000). The metamorphic evolution of the GHS is often split into two tectonothermal events that may mark distinct thermal pulses or a thermal continuum. Kyanite-grade assemblages are interpreted as relicts of an early event (M1) that U Th Pb monazite geochronology suggests occurred at c Ma (Walker et al. 1999; Simpson et al. 2000). M1 is often overprinted by the pervasive high temperature low pressure tectonothermal event (M2; c Ma) associated with decompression, migmatization, and emplacement of leucogranite sills (Hodges 2000; Simpson et al. 2000). 40 Ar/ 39 Ar thermochronology of muscovite and biotite from leucogranites yields cooling ages that are usually only slightly younger than the U Pb crystallization age of the leucogranites, suggesting rapid decompression following their emplacement (e.g. Hodges et al. (1998) for the Everest transect). Two shear zones bound the GHS: the STDS above and the MCTZ below (Figs 1 & 2). The STDS juxtaposes Tethyan sedimentary rocks of the Tibetan Zone against the GHS, while the MCTZ separates the GHS above from rocks of the Lesser Himalaya Zone (LHZ) below (for detailed review see Godin et al. 2006b). Extrusion models Despite on-going controversy regarding evidence for simultaneous movement on the MCTZ and STDS (see Godin et al. 2006b), many researchers continue to view the tectonic evolution of the GHS in the context of models involving southward extrusion of mid-crustal rocks from beneath the Tibetan plateau towards the topographic surface at the plateau margin. All of these models assume simultaneous motion on the upper and lower

4 382 M. J. JESSUP ET AL. surfaces of the extruding unit. Two types of models may be distinguished: (1) kinematic models for wedge extrusion (Fig. 2a c) based on the assumption that the STDS and MCTZ join at depth as, for example, suggested by early interpretations of INDEPTH seismic data (Nelson et al. 1996); (2) more complex coupled thermal mechanical finite-element models involving lateral flow of relatively low-viscosity material within a tabular mid-crustal channel in response to a horizontal gradient in lithostatic pressure between the Tibetan plateau and Himalayan foreland (Fig. 2d; Beaumont et al. 2001, 2004, 2006). Kinematic models Model 1 In the original wedge extrusion model (Burchfiel & oyden 1985; oyden & Burchfiel 1987; Kündig 1988, 1989; Burchfiel et al. 1992; Hodges et al. 1993), the extruding wedge developed by gravity-driven collapse in response to the extreme topographic gradient developed along the southern margin of the Himalayan orogen. In this basic conceptual framework, only one main phase of north south extension, triggered by a fundamental, non-reversible change in the stress state of the orogenic system as a whole (e.g. England & Molnar 1993), occurred in the Himalaya. Synconvergence extrusion, rather than gravitydriven collapse, has been emphasized in more recent models. In the original wedge extrusion model the nature of deformation/flow within the interior portions of the wedge was not explicitly addressed (Fig. 2a), although a broad zone of reverse-sense shearing along the base of the wedge was suggested (e.g. Brunel & Kienast 1986; Hubbard 1988, 1989, 1996; but see also Harrison et al. 1999) as a potential explanation for the long-known inversion of metamorphic isograds adjacent to the MCT (Heim & Gansser 1939). Based on the mapping of antiformally folded isograds in the Zanskar section of the GHS, this was expanded upon in an alternative model by Searle & ex (1989) who proposed that the entire sequence of rocks contained between the MCT and STDS (Zanskar shear zone) was isoclinally folded during extrusion. More recently proposed kinematic extrusion models are largely based on local evidence for spatial strain path (or vorticity) partitioning within the GHS, and the relative importance of pure shear and simple shear flow components. Model 2 Quantitative evidence for a significant pure shear deformation component associated with flow along the base of the GHS was reported by Grasemann et al. (1999) from the Sutlej Valley (NW India) section of the MCTZ (Fig. 1). Based on correlation between: (1) a downward increase in estimated pure shear component, (2) a downward decrease in deformation temperatures within this zone of inverted isograds, and (3) a high pure shear component indicated by late vein sets, Grasemann et al. (1999) proposed that their vorticity data were most readily interpreted as indicating a temporal (rather than spatial) change in flow regime associated with a decelerating strain path (Simpson & De Paor 1997; Fossen & Tikoff 1997), where simple shear flow at higher temperatures is replaced by pure-shear-dominated flow at lower temperatures. Grasemann et al. (1999) proposed a model for wedge extrusion in which deformation is concentrated towards the boundaries of the wedge and, due to strain compatibility, the centre of the wedge extrudes mainly by pure shearing (Fig. 2b). An important aspect of this model is that the wedge is detached from the footwall and hanging wall. Model 3 Microstructural and quartz petrofabric data were employed by Grujic et al. (1996) to qualitatively investigate deformation/flow within the lower-central portion of the GHS exposed in Bhutan. These data indicated that at least the lower-central part of the slab had undergone plane strain to weakly constrictional deformation, with flow involving components of both simple shear (reverse or top-to-the-south shear sense) and pure shear. Based on these data, Grujic et al. (1996) proposed a wedge extrusion model in which pervasive shearing occurred throughout the evolving wedge, with opposite shear senses on the top and bottom halves of the wedge (and highest extrusion velocities in the centre) leading, as previously proposed by Searle & ex (1989) for the Zanskar Himalaya, to antiformal folding of isograds (Fig. 2c; see Godin et al. 2006b). The pure shear component indicated by petrofabric data was not explicitly addressed in this channel flow model, although Grujic et al. (1996) did note that pure shear would lead to thinning and transport-parallel stretching of the wedge/channel during extrusion. Subsequent fieldwork in Bhutan led Grujic et al. (2002) to abandon their wedge-shaped model for the GHS and to regard the GHS as a km thick mid-crustal layer, or channel, extending for at least 200 km northward beneath Tibet. In this revised channel flow model (incorporating combined Couette and Poiseuille flow; see review by Grujic 2006) the influence of changing thermal conditions (viscosity) on flow patterns was considered, and qualitative predictions made on the likely influence of changes in boundary conditions and viscosities on domainal variation in flow vorticities

5 Fig. 3. Simplified geological cross-section of the Mount Everest massif, based on a compilation of original data for this investigation, (Arita 1983; Hubbard 1988, 1989; Lombardo et al. 1993; Carosi et al. 1999a; Catlos et al. 2002; Searle et al. 2003). Metamorphic isograd distribution proposed on the Nuptse Lhotse wall is based on Jessup et al. (2004, 2005) and Waters et al. (2006). Location of key outcrop of folding relationships at base of Lhotse wall is indicated. MCT, Main Central thrust (MCT at top, MCT1 at base); LD, Lhotse detachment; QD Qomolangma detachment.

6 FLOW PATITIONING IN THE HIMALAYA 383 within the channel. Grujic et al. (2002, p. 188) emphasized that general flow (i.e. combined simple and pure shear) is implicit in the Poiseuille flow, and therefore in channel flow. Thermal mechanical models Model 4 Thermal mechanical models build on the concept originally proposed by Nelson et al. (1996) that the GHS represents hot low-viscosity mid-crustal material extruded southwards from beneath Tibet towards the Himalayan front during continental convergence, and the overlapping proposal that this extrusion can be modelled using the concept of channel flow driven by a horizontal gradient in lithostatic pressure between the Tibetan plateau and the Himalayan front (Grujic et al. 1996). These concepts, and a broad range of Himalayan structural, pressure temperature time (P T t) and geochronologic data, have been successfully modelled in two dimensions with timevarying, plane strain, coupled thermal mechanical finite-element models in which channel viscosities are reduced by mantle heat flux and radiogenic heating (Beaumont et al. 2001, 2004, 2006; Jamieson et al. 2002, 2004, 2006). Models begin with a tectonically thickened crust, which is then thermally weakened, and flows in a mid-crustal channel towards the orogenic front. Varying input parameters and model specifications produce variants of the basic model. In these models, channels are exhumed and exposed by denudation focused on the high-relief transition between the plateau and orogenic front (Fig. 2d). Implicit in these models is that the structures now exposed at the topographic front will probably have formed during the last stages, or cessation of extrusion/ exhumation of the channel material, rather than being directly related to processes operating when these rocks were flowing at mid-crustal levels beneath the plateau. Everest transect: geological background In this section we outline the major lithotectonic units and structures of the Everest transect in order to provide a foundation for the detailed discussions of key areas used for our vorticity analyses. This transect begins with the STDS at the top of the GHS (as exposed in ongbuk valley, Tibet) extends southward to the summit of Everest, and continues southward through the GHS to the more limited exposures in the Nepalese foothills of rocks belonging to the MCTZ. Detailed structural, metamorphic and geochronological reviews of the Everest transect (some limited to specific sections) are given by Lombard (1958), Bordet (1961), Brunel & Kienast (1986), Hubbard (1988, 1989), Lombardo et al. (1993), Pognante & Benna (1993), Carosi et al. (1998, 1999a, b), Searle (1999a, b) and Searle et al. (2003, 2006). Geological maps covering different sections of the transect have been published by Bordet (1961), Lombardo et al. (1993), Carosi et al. (1998) and Searle (2003). Two major detachments belonging to the STDS have been mapped in the sidewalls of the ongbuk valley southward to Changtse and Mount Everest: the upper brittle Qomolangma detachment (QD) and lower ductile Lhotse detachment (LD) (Fig. 3; Searle 1999a; Searle et al. 2003; see also Lombardo et al. 1993; Carosi et al. 1998, 1999b; Sakai et al. 2005). As originally defined, the LD is a distinct ductile high-strain zone that marks a metamorphic break between amphibolitefacies rocks below and greenschist-facies (Everest Series) rocks above (Searle 1999a). More recent detailed thermobarometric results from samples collected at the base of the Lhotse wall (Jessup et al. 2004, 2005) and East ongbuk glacier (Waters et al. 2006), demonstrate temperatures of c. 6508C immediately above and below the proposed detachment, and suggest that the LD may mark the upper limit of leucogranites, but not a break in metamorphic grade. The two detachments merge into one major ductile brittle shear zone near the northern limit of ongbuk valley (Fig. 3; Carosi et al. 1998, 1999b; Searle 1999a). Because the QD dips more steeply than the LD, a northward-tapering wedge of Everest Series is mapped between the two detachments (Searle et al. 2002, 2003; Searle 2003). Previous detailed kinematic investigations are limited to the ongbuk valley located on the north side of the Everest massif (Fig. 3). Cross-girdle quartz c-axis fabrics from GHS rocks exposed in the ongbuk valley demonstrate that penetrative deformation, along at least this local section of the STDS, occurred under approximately plane strain conditions, and their asymmetry confirms the top-to-the-north shear sense (Law et al. 2004). Vorticity analysis (using three techniques) on reconnaissance samples, collected from the top of the GHS in the ongbuk area, indicate pure shear components representing c % of the total recorded deformation, depending on rock type and structural position (Law et al. 2004). Integration of strain and vorticity data, in the reconnaissance samples, indicated a shortening of 10 30% perpendicular to the upper surface of the GHS and, as previously suggested by Grasemann et al. (1999) for the MCTZ at the base of the slab in NW India (see below), confirmed that the STDS is a stretching fault (in the sense of Means 1989) with estimated

7 384 M. J. JESSUP ET AL. down-dip stretches of 10 40% (assuming plane strain deformation as demonstrated by petrofabric results) parallel to the flow plane-transport direction. A c. 100 m thick mylonite zone, capped by a breccia zone of variable thickness, characterizes the uppermost section of the GHS in ongbuk valley and projects c. 35 km southwards to the summit of Mount Everest. Structure contours of the detachment (QD) on the summit of Everest, and two peaks to the north (Changtse, 7583 m, and Chang Zheng, 7583 m), suggest the detachment dips c.108 NNE on the summit and shallows to c.58 NNE in the northern limits of ongbuk valley (Fig. 3). On Mount Everest and Changtse, the QD separates Tethyan limestone of presumed early middle Ordovician age (Yin & Kuo 1978) above from underlying marble of the Yellow Band (Everest Series) (Burchfiel et al. 1992; Searle 1999a, 2003). On the NE ridge of Everest, the QD is marked by a 5 40 cm thick breccia zone in the basal limestone, which rests on intensely foliated Yellow Band marble containing shear bands and drag folds (Sakai et al. 2005). The structurally highest section of the Everest Lhotse massif is predominantly composed of greenschist to lower-amphibolite facies Everest Series metasedimentary rocks, while the lower ramparts consist of sillimanite-grade schist that grades into migmatitic gneiss (Fig. 3; Lombardo et al. 1993; Pognante & Benna 1993; Carosi et al. 1998, 1999b; Searle 1999a, b; Searle et al. 2003). A variably deformed leucogranite sill complex that parallels the pervasive fabric within these rocks is limited to a zone immediately below the Everest Series. Searle (1999a) proposed that the LD is present along this transition and also proposed that a late-stage thrust (Khumbu thrust) is present along the base of the underlying, most extensive, leucogranite sill complex (Fig. 3). Variably migmatized, interlayered gneiss, calc-silicate, quartzite, schist and orthogneiss are predominant beneath the LD (or the composite LD QD in the northern ongbuk valley), and extend downwards through the middle section of the GHS to the upper section of the MCTZ (Lombardo et al. 1993; Searle et al. 2003). Deformation within the core of the GHS is characterized by several phases of folding that culminate in a pervasive foliation that is broadly warped by late-stage NWand NE-trending hinge lines of recumbent folds that create dome structures (Carosi et al. 1999a, b). Mylonite zones, typically found at the margins of the slab, are absent in the core. As exposed in the Duhd Kosi drainage south of Everest, the MCTZ consists of sheared quartzite, calc-silicate, amphibolite, garnet kyanite staurolite schist, graphitic schist and augen gneiss (Hubbard 1988, 1989; Catlos et al. 2002). Although debate continues about the exact location of the thrust zone (see Godin et al. 2006b; Searle et al. 2006), we choose to relate our vorticity results to the original context proposed by Hubbard (1988, 1989). In the Duhd Kosi drainage, a combined downward increase in apparent penetrative strain, and upward increase in metamorphic temperatures that exceed the kyanite stability field, marks the top (MCT) of the 5 km thick high-strain zone, while the Okhandunga orthogneiss marks the base (MCT I; Fig. 4). Because the apparent increase in strain coincides with a change in rock type (migmatitic gneiss above and pelitic schist below), the downward increase in penetrative foliation intensity may be controlled by lithology rather than structural position. The pervasive north-dipping foliation overprints several phases of folding and foliation development that are only preserved in lower- to moderate-strain domains S GANET STAUOLITE KYANITE SILLIMANITE 87-H-5A MCT I 87-H-1B 87-H-6B garnet mica schist graphitic schist garnet mica schist 87-H-22E ET H-21J 87-H-21G amphibolite ET-44 augen gneiss marble quartzite MCT biotite gneiss N 4 3 elevation (km) above sea level Okhandunga gneiss Phaplu augen gneiss 2 Fig. 4. Generalized cross-section of the Main Central thrust zone (MCTZ), after Hubbard (1988). Locations of isograds are approximate. Dip of the units is based on work from this investigation.

8 FLOW PATITIONING IN THE HIMALAYA 385 within the high-strain zone. Variation in foliation orientation is the result of late-stage folding also present at structurally higher positions in the core of the GHS. Vorticity analysis Introduction to techniques Mean kinematic vorticity number (Wm) is a measure of the relative contributions of pure (Wm ¼ 0) and simple (Wm ¼ 1) shear. Several analytical methods exist for estimating Wm in high-strain rocks; however, only in rare cases are individual samples suited for vorticity analysis using multiple techniques (see Law et al. (2004) for detailed discussion). We focus on a suite of samples collected from the Everest transect that are suitable for the rigid grain-based vorticity analysis developed by Wallis et al. (1993) and Wallis (1995). This suite of 51 samples (Table 1) includes the seven reconnaissance samples described by Law et al. (2004) from the ongbuk valley Changtse ridge part of the transect, some of which had previously proved suitable for multiple methods of vorticity analysis. Using the founding principles of Ghosh & amberg (1976) and Passchier (1987), Wallis et al. (1993) and Wallis (1995) proposed that, for rigid clasts rotating in a flowing ductile matrix, a unique relationship exists between Wm, clast aspect ratio () and the angle (u) between clast long axes and matrix foliation. For a given Wm, clasts with a specific aspect ratio will reach a unique stable sink position (i.e. angle from the foliation). The method involves measuring the clast aspect ratio and angle between the clast long axis and foliation for both back- and forwardrotated clasts (in sections cut perpendicular to the foliation and parallel to the macroscopic stretching lineation). The distribution of clasts is displayed on a plot of versus u (Fig. 5a d). A transition between clasts that rotate infinitely and those that reach a stable sink orientation defines the critical threshold ( c ). c is then used to calculate Wm using the relationship proposed by Passchier (1987): Wm ¼ð 2 c 1Þ=ð2 c þ 1Þ (1) In practice, a range of likely c values is usually indicated for a given sample using the Wallis plot, leading to a range of estimated Wm values (Law et al. 2004; see also Carosi et al. 2006; Xypolias & Kokkalas 2006). Whether the Wallis method may consistently under- or overestimate Wm values probably depends on individual sample characteristics. The method may tend to underestimate Wm if clasts of large aspect ratio are not present, and in such samples the upper bound of the estimated Wm range is probably closest to the true value (Law et al. 2004). In contrast, if finite strains are low, then clasts of high aspect ratio may not have had time to reach stable sink orientations and the observed range of c values would tend to overestimate Wm (Bailey et al. in press). The rigid grain vorticity method assumes: (1) that the clasts undergo no internal deformation (e.g. by crystal plasticity or pressure solution); (2) no mechanical interaction occurs either between adjacent rotating clasts, or between the clasts and their matrix; and (3) high enough strain has developed to ensure that all clasts have rotated into their current position. Samples were avoided where deformation temperatures exceeded the onset of internal plastic deformation within otherwise rigid phases, or where excessive interaction had occurred between rotated grains. We tentatively assume plane strain deformation for these samples, based on the original petrofabric data of Law et al. (2004, p. 313), which strongly indicated that flow was monoclinic to orthorhombic (Law et al. 2004, p. 314). Due to the lack of robust strain markers, it was impossible to quantify strain in any of the samples from the Everest transect aside from those published by Law et al. (2004). igid grain data plots for all 51 samples used for vorticity analysis are reproduced in the Appendix to this paper. Full details of the mineral(s) used as rigid grain markers are given on each plot. Details of sample locations, and estimated range of Wm values for each sample, are summarized in Table 1. epresentative rigid grain data plots for different structural levels Four representative samples are used to describe and discuss the characteristics of the major rock types within the Everest transect used for vorticity analysis: (1) sheared Tethyan limestone above the QD system; (2) biotite gneiss/schist within the uppermost 100 m of the footwall to the composite QD LD system; (3) high-grade gneiss at a deeper structural level (,2 km) in the footwall to the LD; and (4) pelitic rocks within the MCTZ (Fig. 5a d). Tethyan limestone Sample GB-25/3, collected from just below the summit (8836 m) by Edmund Hillary (Harker Collection records, Cambridge University) during the first accent of Mount Everest via the South Col in 1953, is an example

9 386 M. J. JESSUP ET AL. Table 1. Mean kinematic vorticity (Wm) data Sample ock type Elevation (m) Distance (m) from QD LD Method 1 (Wm) Northern transect lim above mar calc calc qtz calc (A) calc calc leu leu leu leu leu (A) bt ongbuk Monastery transect lim above lim above lim/mar mar calc TI-05 bt c ET-15 bt calc ET-14 bt bt ET-13 bt ET-12 bt Hermit s Gorge transect lim above mar mar calc bt ET-08 þ bt leu bt Everest & Kangshung valley transects 25/3 Hillary lim E Hamilton lim ME-124 Wager lim /1&2 Evans lim ME-125 Wager lim ET-10 bt ET-11 bt K04-03 gn K04-04 gn Main Central Thrust zone transect ET-41 grt-sill ET-44 grt-sill H-22E grt-sill H-21J grt-ky H-21G grt-ky H-6B grt H-5A gn H-1B grt bt, biotite schist; calc, calc-silicate; gn, gneiss; grt, garnet schist; grt-ky, garnet þ kyanite schist; grt-sill, garnet þ sillimanite schist; leu, leucogranite; lim, limestone; mar, marble; qtz, quartz-rich layer in calc-silicate.

10 FLOW PATITIONING IN THE HIMALAYA 387 angle between clast long axis and foliation angle between clast long axis and foliation angle between clast long axis and foliation angle between clast long axis and foliation c = 2.10 c = 2.65 c = (b) Wm = n = 200 feldspar tourmaline c = 2.55 c = 3.80 c = (a) GB-25/3 - Hillary Wm = n = 208 c = 2.30 c = 2.80 feldspar tourmaline quartz clast in calcite matrix (c) K Wm = n = 200 (d) ET-41 Wm = n = 227 garnet tourmaline Fig. 5. igid grain plots using the Wallis et al. (1993) and Wallis (1995) technique (see text for details). Four representative plots are used to discuss the main rock types used in this investigation: (a) sheared limestone in the hanging wall of the Qomolangma detachment; (b) mylonitic metapelites, calc-silicates and leucogranites from the upper 600 m of the GHS; (c) GHS gneiss sample from Kangshung valley with broad range of potential c values which are typical of samples where the originally rigid phase (feldspar) begins to deform internally; (d) garnet schist from the Main Central thrust zone where rigid elongate garnets were used to estimate Wm. of sheared limestone in the immediate hanging wall to the QD (Fig. 3). Other samples that share the same microstructural characteristics and spatial proximity to the QD, were collected in the sidewalls of the ongbuk valley. Abundant equant elongate detrital quartz grains are interpreted as rigid clasts that rotated within a ductile calcite matrix (Fig. 6a). The distribution of quartz grains on the Wallis plot defines (Fig. 5a) an abrupt transition from grains that rotate infinitely ( 3.80) to those that reach a stable sink orientation ( 4.05). Using this range in c values ( ) yields a Wm estimate of (c % pure shear). Immediate footwall to LD and composite QD LD system Sample represents the amphibolitefacies rocks (marble, calc-silicate, leucogranite and biotite schist/gneiss) within the upper 100 m of the GHS. These samples often contain several rigid phases such as feldspar, epidote, zircon, amphibole and tourmaline in a matrix of dynamically recrystallized quartz. At the upper limit to these deformation temperatures (amphibolite facies), large feldspar porphyroclasts remain rigid while smaller grains begin to deform internally is a biotite schist with abundant feldspar and tourmaline suitable for rigid grain analysis (Fig. 6b). The narrow range in c (Fig. 5b) yields a fairly robust Wm estimate of (45 42% pure shear). Structurally deeper levels of LD footwall Sample K was collected from outcrops of highgrade gneiss in the western end of the Kangshung valley. These gneisses are situated at a structural depth of c. 2 km beneath the LD (Fig. 3) and continue downward into the underlying anatectic core of the GHS. Feldspar grains in these gneisses are generally separated from each other by a matrix of biotite laths and dynamically recrystallized (egime 3 of Hirth & Tullis 1992) quartz (Fig. 7a). However, many of the feldspar grains exhibit at least moderate undulatory extinction and minor grain flattening, indicating that they did not behave as perfectly rigid markers. igid grain plots using these feldspar grains are characterized by a broad transition in potential c values (Fig. 5c), and therefore greater uncertainty in defining Wm. We propose that these plots are typical of samples that contain a semi-rigid phase, and caution against overinterpretation of Wm estimates from such samples. MCTZ The fourth example, sample ET-41, is a garnet mica schist typical of pelite samples collected from the MCTZ. These pelite samples contain elongate garnet porphyroclasts that are wrapped by biotite and muscovite, and surrounded by a matrix of quartz and feldspar (Fig. 7b). The

11 388 M. J. JESSUP ET AL. Fig. 6. (a) Photomicrograph (crossed polars) of sample GB-25/3 (collected by E. Hillary in 1953 at c m) showing microstructures typical of sheared limestone collected near the Qomolangma detachment. Abundant detrital quartz grains act as the rigid phase rotating in a calcite (Cal) matrix. Some randomly orientated white mica (M) is present. igid grain plot of the sample is shown in Figure 5a. (b) Photomicrograph (crossed polars) of sample 03-38; section cut perpendicular to the foliation and parallel to the lineation. Microstructures include rigid feldspar (Fs) rotating in a ductile quartz (Qtz) matrix. Large feldspar porphyroclasts in the centre of the image have an aspect ratio of c. 1.6 with a long axis c.808 from the foliation as defined by aligned white mica (M). igid grain plot of the sample is shown in Figure 5b. Sample cut perpendicular to foliation and parallel to lineation; plunge and trend indicated.

12 FLOW PATITIONING IN THE HIMALAYA 389 evolution of these elongate garnets is discussed in detail below. The orientation distribution and range in aspect ratio of these garnets confirms their appropriateness for rigid grain vorticity analysis (Fig. 5d). A limited range in c defines Wm estimates of (58 48% pure shear). The major drawback to using metamorphic phases for rigid grain analysis in these pelitic MCTZ samples is the number of appropriate porphyroblasts available within a thin section; where possible we have used combined data from parallel sections in individual samples. For example, sample 85-H-21G contained a minimal number of garnet porphyroblasts (n ¼ 59) that just begins to define a minimum c, whereas 87-H-22E contained many garnets (n ¼ 275) that define c much better (Appendix, Sheet 5). For several MCTZ samples, such as ET-41, the rigid grain analysis proved highly successful and provides a unique opportunity to explore the relationship between peak metamorphism and mylonite formation. Petrography and results of vorticity analyses ongbuk valley transects We collected orientated samples for vorticity analysis along three transects in the eastern sidewalls of the ongbuk valley (Fig. 8). A similar lithotectonic sequence is observed in each transect consisting (traced structurally downwards) of limestone, marble, calc-silicate, leucogranite and biotite sillimanite schist/gneiss (Figs 8 & 9). Tethyan limestone forms the structurally highest lithotectonic unit and is truncated along the base by the underlying QD. A 5 10 m thick section of marble marks the upper limit to pervasive ductile deformation beneath the detachment. Lenses of mylonitic leucogranite are commonly found within the sheared marble, demonstrating that ductile deformation outlasted their emplacement (c. 17 Ma; Murphy & Harrison 1999). Interlayered and pervasively foliated calc-silicate and quartzofeldspathic layers, defined in outcrop by alternating black/green and white layers, are present below the sheared marble. Dark layers contain diopside and are either amphibole- or tourmaline-rich, while white layers contain abundant quartz and feldspar. Feldspar is commonly fractured and within one thin section a complete gradation from angular clasts to rounded porphyroclasts rotating in a quartz matrix is common. Microstructures in quartz-rich layers include the development of subgrains and bulging grain boundaries, which indicate dynamic recrystallization under egime 2 3 conditions as defined by Hirth & Tullis (1992), and suggest deformation temperatures of c C (Stipp et al. 2002). Microboudinage of diopside, garnet and tourmaline grains suggests that some components of fabric development post-dated their growth. Tension gashes, nearly perpendicular to the NNE- or SSW-trending stretching lineation, suggest that a progression in deformation mechanisms from ductile to brittle occurred during exhumation of the GHS. Structurally beneath the calc-silicate layers (at least at the northern end of the ongbuk valley) is a m thick mylonitic leucogranite sill complex. Quartz and feldspar record evidence for grain-scale processes operating at similar deformation conditions to those indicated in the overlying calc-silicate-rich unit. S-C fabrics with extensional shear bands dominate the detachmentparallel sills. The structurally lowest unit exposed in the ongbuk valley is composed of biotite schist (ongbuk Formation of Carosi et al. 1998, 1999a) that is migmatized and injected by foliation-parallel, variably deformed, leucogranite lenses and sills; cross-cutting leucogranites are less commonly observed (Searle et al. 2006, figs 6 & 7). Based on quartz c-axis fabric opening angles, Law et al. (2004) documented progressive increasing deformation temperatures of C in the biotite schists at depths of m beneath the mapped position of the LD in the ongbuk Monastery and Hermit s Gorge transects (see below). otation of the rigid grains used as vorticity markers in this paper, either pre-dated or (more likely) were synchronous with plastic flow of the quartz-rich matrix associated with these deformation temperatures. Fibrolite in the biotite schist is drawn into extensional shear bands, but remains pristine, suggesting that shear band development occurred in the sillimanite stability field (Law et al. 2004). At depths greater than 100 m beneath the composite QD LD system, feldspar begins to deform plastically (as indicated by undulose extinction) and grains tend to become more elongate and orientated subparallel to foliation. At a given depth, this brittle plastic transition in feldspar deformation seems to be grainsize controlled (Law et al. 2004, p. 311). Incipient conjugate sets of shear bands, defined by biotite, create a lattice network that dominates the microstructure in the structurally deeper samples. Polygonal quartz grains are common, suggesting a component of annealing. Many of these structurally deeper samples are unsuited for vorticity analysis (as discussed above for sample K-04-03). However, even at depths of 600 m beneath the detachment, samples with limited evidence for internal deformation of feldspar yield a

13 390 M. J. JESSUP ET AL. Fig. 7. (a) Photomicrograph (crossed polars) of sample K04-03 collected in Kangshung valley, Tibet. Biotite (Bt) defines the foliation that is aligned NW SE in the image. Irregularly shaped feldspar (Fs) that begins to align with the foliation suggests high deformation temperatures where feldspar begins to deform internally. igid grain plot of this sample is shown in Figure 5c. These microstructures typify samples from the core of the Greater Himalayan Slab that are unsuited for rigid grain analysis. (b) Photomicrograph (crossed polars) of sample ET-41 from the Main Central thrust zone (Fig. 4). Garnets (Grt) of variable aspect ratios and angles from the foliation are present. Quartz (Qtz) inclusions are present in several of the garnet cores. Aligned biotite (Bt) and white mica (M) define the foliation (east west in image). igid grain plot using garnet porphyroblasts is shown in Figure 5d. Details of garnet evolution are discussed in the text. Sample cut perpendicular to foliation and parallel to lineation; plunge and trend indicated.

14 FLOW PATITIONING IN THE HIMALAYA 391 Fig. 8. Simplified geological map of ongbuk valley, Tibet. Insets (a c) are enlargements of detailed sample transects. Spatial distribution of samples is also shown on the cross-section through each transect. North is oblique to the long axis of the figure. Image compilation created using original mapping from this investigation and other sources (Burchfield et al. 1992; Murphy & Harrison 1999; Searle et al. 2003; Law et al. 2004).

15 392 M. J. JESSUP ET AL. Fig. 9. Photograph of the composite Qomolangma and Lhotse detachments (black line) where they are proposed to merge in the northern limits of ongbuk valley. Location where image was taken is shown as solid black star on Figure 8a. The northern transect is located where the road and detachment are closest. The general rock types from structurally highest to lowest are: (1) limestone, (2) marble, (3) calc-silicate, (4) leucogranite, and (5) migmatized biotite schist (also known as ongbuk Formation). View is towards the NE. An arrow points to jeep on two-lane dirt road for scale. well-defined c threshold, and therefore a meaningful Wm estimate. Below we summarize the results of vorticity analyses in our three transects through the eastern sidewalls of the ongbuk valley (Fig. 8); each transect begins in the sheared limestone or within the composite QD LD system and progresses downwards into the migmatitic biotite schist. esults are presented on plots of Wm versus relative distance below the QD to show the spatial distribution of Wm domains in each transect (Figs 10 & 11). Because the location of the QD is more readily determined in the field than the LD, and in the northern section of ongbuk valley the LD either merges with or is cut out by the QD, we use the QD as a reference structural level in these plots. Northern transect Vorticity analysis results from the northern transect (Fig. 8 inset a & Fig. 9) are shown in Figure 10a. Sample (limestone) and sample (marble), collected c. 10 m above and within the QD, respectively, yield Wm estimates of and , indicating the lowest component of pure shear (30 25%) for the entire transect. Four out of six calc-silicate samples from below the sheared footwall marble yield a range in Wm of (c % pure shear). The other two calc-silicate samples are outliers to this trend and yield slightly higher (03-17: Wm ¼ , 40 35% pure shear) and lower (03-19: Wm ¼ , c % pure shear) Wm estimates. The large range in potential c values recorded by calc-silicate samples and 19 (together with leucogranite sample 03-20) suggests they are less suitable for rigid grain analysis than the other samples. Five leucogranite samples yield a range in Wm estimates that is consistent with the majority of the calc-silicate samples ( ). Although large, the range in Wm for sample overlaps with Wm values in the calc-silicate and leucogranite samples. The single biotite schist sample (03-26A) at the base of the transect has a narrow range in estimated Wm values ( ) that is indistinguishable from the calc-silicate and leucogranite samples. The sheared limestone and marble in the immediate hanging wall and footwall to the QD yield the highest Wm values ( ), and therefore highest percentage simple shear values, recorded in the ongbuk valley transects. At distances of c m beneath the detachment, the majority of samples yield Wm estimates of (c % pure shear). ongbuk Monastery transect The ongbuk Monastery transect is located c. 7 km to the south of the northern transect (Fig. 8, inset b). Three limestone samples at the top of the transect (03-55, 56 and 58) yield Wm estimates of , with the greatest simple shear component (c. 63%) recorded in the structurally highest sample (Fig. 10b). The one marble sample (03-59), located beneath the detachment, yields a Wm estimate ( ) that is indistinguishable from Wm values for the

16 FLOW PATITIONING IN THE HIMALAYA 393 (a) NOTHEN TANSECT percent pure shear L DETACHMENT M Altitude (m) (b) ONGBUK MONASTEY percent pure shear L L Altitude (m) C C L DETACHMENT M C C A C C TI-5 ET-15 A G ~ 5600 (talus) 5450 C C 5380 G ET-14 B G B G 4974 ET-13 B G 4965 ET-12 B A B G mean vorticity number - Wm igid grain technique Vorticity method II (Law et al. 2004) Vorticity method III (Law et al. 2004) L: limestone M: marble C: calc-silicate G: leucogranite A: amphibole schist B: biotite gneiss/schist 0.9 mean vorticity number - Wm Fig. 10. Bar charts for range of mean kinematic vorticity numbers (Wm) estimated by the rigid grain method for samples collected in the Northern (a) and ongbuk Monastery (b) transects, ongbuk valley, Tibet. Sample locations shown in Figure 8. The range of Wm values estimated by alternative methods (Law et al. 2004) is also indicated. limestone sample above and the calc-silicate sample below (03-63: Wm ¼ ). Five of the seven samples below calc-silicate 03-63, including one calc-silicate (03-67), one hornblende epidote schist (TI-5), and three biotite schist samples, record a fairly consistent range in estimated Wm values ( ). The two outliers yield slightly higher (leucogranite ET-15: Wm ¼ ) and lower (biotite schist ET-14: Wm ¼ ) Wm estimates. Samples TI-5, ET-14, ET-13 and ET-12 also proved appropriate for several other vorticity analysis techniques (Law et al. 2004), referred to in Figure 10b as method II (the PHD method of Simpson & De Paor 1997) and method III (the combined strain and quartz c-axis fabric method of Wallis 1995). For TI-5, the rigid grain

17 394 M. J. JESSUP ET AL. Fig. 11. Bar charts for range of mean kinematic vorticity numbers (Wm) estimated by the rigid grain method for samples collected in the Hermit s Gorge (a) and Mount Everest & Kangshung valley (b) transects, Tibet. Sample locations shown on Figures 8, 12 & 13. The range of Wm values estimated by alternative methods (Law et al. 2004) is also indicated. technique of Wallis et al. (1993) and method II yield indistinguishable results. For the other three samples, method III consistently yields higher Wm estimates than the rigid grain technique (see Law et al. (2004) for detailed discussion). In summary, rigid grain analyses from the ongbuk Monastery transect yield Wm estimates of (c % pure shear) and represent deformation conditions to a maximum depth of c. 600 m beneath the composite QD LD fault system. We regard the structurally deepest samples as yielding the least reliable Wm estimates, as all size fractions of feldspar grains display at least limited evidence for crystal plasticity, and thereby undermine the fundamental assumptions of the rigid grain technique. Hermit s Gorge transect Our third transect is located in Hermit s Gorge (and one of its side valleys) which intersects the ongbuk valley at Everest Base Camp (Fig. 8, inset c). One sheared limestone sample (03-46) was collected c. 5m above the top of the marble section and presumed location of the QD. It yields a narrow range in Wm estimates of (c. 45% pure shear) that is indistinguishable from the two marble samples (03-43 and 44) below. The single calcsilicate sample (03-39) yields a Wm estimate

18 FLOW PATITIONING IN THE HIMALAYA 395 Fig. 12. Simplified geological map of the Mount Everest massif and Kangshung valley, Tibet. Compilation based on mapping during this project (Kangshung valley) and Searle et al. (2003). QD, Qomolangma detachment; LD, Lhotse detachment. ( ) that is significantly lower than both the marble above and biotite schist below (03-38: ). Sample ET-8, a biotite-rich psammite, yields the highest Wm estimate of the entire transect ( ); in contrast, method III analysis on this sample yields higher estimated Wm values (Law et al. 2004), as noted for samples from the ongbuk Monastery traverse , a piece of mylonitic leucogranite float collected at an altitude of c m, yields the lowest Wm estimate of Although collected at the highest altitude of the transect, due to its position on the south side of the gorge and the northerly dip of the structural units, this sample probably comes from a relatively deep structural position. The structurally lowest sample (03-33), collected near the mouth of Hermit s Gorge at c. 340 m beneath the QD, yields the largest range in estimated Wm values ( ) for the transect. We attribute the large range in uncertainty of c (and hence Wm) for this sample to the onset of plastic deformation in the feldspar marker grains. This sample is probably close to the maximum structural depth for robust rigid grain vorticity analysis. Summit of Mount Everest Kangshung valley Our final transect across the top of the GHS is composed of Tethyan limestone and Yellow Band (Everest Series) marble samples from near the summit of Mount Everest, samples of Everest Series interlayered pelite and calc-mylonite collected from talus piles at Advance Base Camp beneath the North Col Changtse idge, and

19 396 M. J. JESSUP ET AL. Fig. 13. Photograph of the summit of Mount Everest (viewed towards the east) taken from enjo La (5340 m), Nepal, using a 300 mm lens. Qomolangma detachment is highlighted by line and separates Tethyan limestone (1) above from Yellow Band marble (2) and Everest Series (3) below. Approximate locations of samples collected by Wager in 1933 (ME-124 and 125), Evans in 1953 (GB-25/1 and 2), Hillary in 1953 (GB-25/3), and Hamilton in 2003 (E-03-01) are indicated. Photomicrograph of summit sample collected by Hillary is shown in Figure 6a, and the corresponding rigid grain plot is shown in Figure 5a. samples of high-grade gneiss from outcrops in the western end of the Kangshung valley (Figs 3, 12 & 13). Only the Kangshung valley samples, discussed above (Fig. 5), are orientated. The highest altitude sample (GB-25/3), from the Harker Collection at Cambridge University, is of Tethyan limestone collected by Edmund Hillary on 29 May 1953 at 40 feet beneath the summit of Mount Everest (Harker Collection records). This sample is augmented by a second, lithologically identical, summit sample (E-03-01) collected by Scottish alpinist David Hamilton in Our third, and structurally deepest, Tethyan limestone sample (ME-124), from the Lawrence Wager Collection in the Oxford University Museum of Natural History, was collected by Wager from a band forming the First Step (Wager Collection records; see also Wager 1934, 1939) on the NE ridge of Everest during the 1933 Everest expedition. The highest altitude Yellow Band marble sample (GB-25/1 þ 2), two pieces of intensely foliated and lineated white calc-mylonite from the Harker Collection, was collected by Charles Evans on 26 May 1953 at approximately feet (Harker Collection records) on the SE ridge of Everest. Our structurally deeper Yellow Band sample (ME-125) was collected by Wager from a typical yellow schistose marble forming Yellow Band on the NE ridge at approximately 300 feet beneath the 1933 Camp VI (Wager Collection records). Tethyan limestone Sheared Tethyan limestone samples (GB-25/3, E03-01, ME-124) contain abundant white mica laths and subangular subrounded detrital quartz grains set in a calcite matrix (Fig. 6a). The calcite matrix grains are completely recrystallized, and no remnants of a sedimentary fabric have been preserved (J.F. ead, pers. comm. 2006). The calcite grains are equant slightly elongate in cross-section, and an incipient foliation is defined by weak preferred orientation of the more elongate matrix grains, together with aligned films of an extremely fine-grained opaque phase. Calcite- and quartz-filled microfaults truncate the incipient foliation at moderate to high angles, particularly in sample ME-124 collected from immediately above the QD (Fig. 13). Anastomosing quartz-filled fractures subparallel to foliation are also present. The matrix calcite grains range in size from 20 to 50 mm. Larger single and polygonal calcite grains ( mm), together with randomly orientated white mica laths (up to 100 mm in length) and equant elongate detrital quartz grains (generally mm long), are scattered throughout the matrix (Fig. 6a). Weak undulose extinction within

20 FLOW PATITIONING IN THE HIMALAYA 397 the quartz grains suggests a minor component of plastic deformation, and a high concentration of fluid inclusions gives a dusty appearance to some of these grains. E-twins in the larger calcite grains are straight and thin (,5 mm), suggesting deformation temperatures, C (Burkard, 1993; Ferrill et al. 2004). The presence of slightly wider twins (.5 mm) in some of the smaller matrix grains suggests that deformation temperatures may have reached.2008c (Burkhard 1993; Ferrill et al. 2004). Observed microstructures, and well-defined c values in all three of these subgreenschist-facies Tethyan limestone samples indicate that the detrital quartz grains acted as at least semi-rigid clasts rotating in a plastically flowing and dynamically recrystallizing calcite matrix. Wm estimates (Fig. 11b) in these samples range from 0.84 to 0.89 (35 30% pure shear). The pristine grain boundaries of the white mica laths in the limestone suggest that they may have recrystallized during deformation, rather than being of detrital origin (G. Oliver pers comm. 2006). We attribute the lack of a well-developed grain-shape foliation within the calcite matrix, together with the lack of any sedimentary structures, to the operation of grain boundary migration recrystallization, as indicated by the observed microstructures in this sample. Samples of Everest summit limestone have previously been described by Gansser (1964, pp ) and Sakai et al. (2005). The microstructures described by Gansser (including samples originally described by Gysin & Lombard 1959, 1960) are very similar to those recorded in our samples, except for the presence of crinoid fragments (see also Odell 1965). In contrast, the sample described by Sakai et al. (2005), and collected at c. 6 m beneath the summit (8850 m), contains crinoid, brachiopod and trilobite fragments, and seems to be much less extensively sheared and recrystallized. Everest Series, Yellow Band marble Microstructurally, the most obvious difference between the summit limestone and the underlying Yellow Band marble is the change in size of recrystallized matrix calcite grains, which abruptly increases from mm in the Tethyan limestone above the QD to mm in the Yellow Band marble beneath the detachment. The calcite grains are equant to slightly elongate and define a weak foliation in thin section that is parallel to the strong macroscopic foliation. Larger single calcite grains ( mm), together with randomly orientated white mica laths (up to 100 mm in length) and equant elongate detrital quartz grains (generally mm long), are scattered throughout the matrix. Calcite twins are thicker and more closely spaced than in the Tethyan limestone, and both multiple twin sets and tight chevron-style buckling of twin lamellae are commonly developed in the larger calcite grains (particularly in sample GB-25/1 þ 2). The presence of thick twins and microstructural evidence for widespread calcite recrystallization involving grain boundary migration indicates deformation temperatures.2508c (Ferrill et al. 2004). However, the detrital quartz grains exhibit very little undulose extinction, and appear to have acted as semi-rigid porphyroclasts in the flowing calcite matrix, indicating deformation temperatures, C (i.e. below generally accepted minimum temperatures for onset of plastic deformation in quartz at natural strain rates; see Stipp et al. (2002) and references therein). Wm values of (36 32% pure shear) are indicated for samples 25/1þ2 and ME-125 using the detrital quartz grains as rigid markers (Fig. 11b). Structurally deeper levels of Everest Series Talus samples of biotite-grade interlayered phyllite psammite and calc-mylonite (ET-10 and ET-11), shed from the structurally deeper sections of the Everest Series exposed on the North Col Changtse idge (Figs 3 & 12), clearly indicate a strong matrix control on deformation mechanisms operating in detrital quartz grains. Even at the thin-section scale, a strong partitioning of deformation mechanisms is observed. Detrital quartz grains deform plastically (with minor pressure solution) in the pelite layers when surrounded by phyllosilicates (biotite and white mica), but remain as rigid clasts in the calc-mylonite layers where the calcite grains have accommodated the penetrative strain. Wm values of (42 37% pure shear) are indicated for samples ET-10 and ET-11 using the detrital quartz grains in the calcite-rich layers as rigid grains (Fig. 11b). The combined strain and quartz c-axisfabricmethodofwallis(1995)inthe quartz mica layers yielded Wm estimates of (Law et al. 2004) and correspondingly lower pure shear components (Fig. 11b; ET-10, method III). Main Central Thrust Zone The base of the GHS is marked by the MCTZ (Fig. 3). In the Everest transect the MCTZ is approximately 5 km thick, and characterized by a general decrease in metamorphic grade towards deeper crustal levels (Fig. 4), as constrained by the appearance of index minerals and geothermobarometry (Hubbard 1988, 1989; Searle et al. 2003). Seven samples of garnet-bearing schist and

21 398 M. J. JESSUP ET AL. Fig. 14. (a) Scanning electron microscope (SEM) image of garnets typical of sample ET-41 from the Main Central thrust zone. Foliation is defined by aligned muscovite (intermediate grey). Biotite (light grey) forms tails on some garnet porphyroblasts and also defines the foliation. (b) SEM image of an elongate garnet in sample ET-41. Foliation is oriented east west in the image. Sigmoidal inclusion trails defined by quartz (dark grey), biotite (intermediate grey), and oxides (white). Inclusion-free rims are preserved on both ends of the garnet. (c) SEM image of another example of an elongate garnet porphyroblast in sample ET-41. (d) Three-step evolution of elongate garnets used for rigid grain analysis (see text for details). igid grain plot for this sample is shown in Figure 5d and a photomicrograph in Figure 7b. Sample cut perpendicular to foliation and parallel to lineation.

22 FLOW PATITIONING IN THE HIMALAYA 399 one sample of orthogneiss were selected from different structural levels of the MCTZ for rigid grain analysis (Figs 3 & 4). Three basic types of garnet grains are distinguished in the schist: round garnets, small irregularly shaped garnets, and elongate garnets (Fig. 14). ound garnets preserve concentric zoning defined by inclusion-rich (commonly sigmoidal) cores and inclusion-free rims (Fig. 14a). Elongate garnets commonly contain sigmoidal inclusion trails or more planar inclusion trails at a high angle to the grain long axis (Fig. 14b, c). Small irregular garnets are dominantly inclusion-free (Hubbard 1988, 1989). We propose a three-step evolution for these garnets (Fig. 14d, steps 1 3): (1) formation of inclusion-rich cores during initial garnet nucleation and growth (as preserved by round garnets); (2) growth of inclusion-free rims during a second phase of garnet growth; (3) local removal of garnet rim/core material by a combination of brittle fracturing and pressure solution during latestage penetrative shearing and foliation development within the MCTZ. At least some of the observed irregular garnets may be fracture fragments. These microstructures (Fig. 14) indicate that deformation associated with both rotation of these elongate truncated garnets, and formation of the observed enveloping penetrative foliation, must either post-date or have outlasted peak metamorphic conditions, as previously suggested by Hubbard (1988, 1989, 1996); see also Brunel & Kienast (1986) for a similar interpretation alongstrike in the Makalu section of the MCTZ. esults of our vorticity analyses, based on the dispersion of these garnet porphyroblasts, must also relate to penetrative flow that outlasted or post-dated peak metamorphism. Many of the garnets in these samples are the same ones used by Hubbard (1988, 1989) to define the inverted metamorphic isograds along the Dudh Kosi section of the MCTZ. Therefore, these isograds may have formed prior to this phase of deformation and shearing along the MCTZ (Hubbard 1996), which is potentially associated with relative late-stage extrusion of the GHS. Locations of samples used for rigid grain analysis are shown in a schematic cross-section through the MCTZ (Fig. 4). The structurally highest samples (ET-44, ET-41 and 87-H-22E) are within the sillimanite stability field of the MCTZ and yield Wm estimates of (c % pure shear; Fig. 15). Elongate garnet porphyroclasts in kyanite-bearing samples (87-H-21J and 87-H- 21G) yield Wm estimates of (c % pure shear). One sample (87-H-6B), thought to roughly coincide with the staurolite zone, yields a similar Wm estimate of (c. 50% pure shear). Sample 87-H-5A, collected 0.5 MAIN CENTAL THUST percent pure shear S igid grain technique S: schist; GN: gneiss ET-44 ET H-22E 85-H-21J 85-H-21G S 87-H-6B MAIN CENTAL THUST I GN 87-H-5A 0.6 S S S S 0.7 S H-1B 0.9 mean vorticity number - Wm from within a sheared section of the Okhandunga gneiss, yields a Wm estimate of using feldspar grains. Sample 87-H-1B was collected further south, in an essentially unmapped section of the MCTZ, yet yields a Wm estimate of using garnet porphyroblasts. The range in Wm estimates from these MCTZ samples ( ; average minimum and maximum Wm values of 0.67 and 0.72) suggests a c % pure shear component at the base of the GHS following peak metamorphic conditions. Core of the Greater Himalayan Slab Vorticity analyses in the anatectic core of the GHS were not possible because these high-grade rocks 10 sillimanite kyanite kyanite staurolite staurolite garnet Fig. 15. Bar chart for range of mean kinematic vorticity numbers (Wm) estimated by the rigid grain method for samples collected in the Main Central thrust zone, Khumbu region, Nepal. Metamorphic isograd locations are approximate. Sample locations shown in Figures 3 & 4.

23 400 M. J. JESSUP ET AL. lack mineral phases that remain rigid at high deformation temperatures. Deformation in the core of the GHS is markedly different from along its bounding margins (STDS and MCTZ). Polyphase deformation of the metasedimentary rocks produced at least two phases of folds that are migmatized to variable degrees and injected by numerous leucogranite sill complexes. Key overprinting relationships exposed throughout the core provide critical insight into the structural evolution of the GHS. One such exposure is located on the lower ramparts of the Nuptse Lhotse wall where at least two phases of folding are preserved in a single outcrop composed of interlayered quartzite and pelites (Figs 3 & 16). The first phase of deformation (D1) is recorded by isoclinal F1 folds (148! 298W) that fold graded bedding in quartzite and create a composite S0 S1 foliation (N158E, 358NW). Within the F1 hinges zones, white mica is aligned at a high angle to S0 and defines an axial planar foliation. Quartz microstructures within the quartzite layers record a limited degree of annealing. Isoclinal folds were refolded during a second phase of deformation (D2), producing both open F2 folds (138! N388W) that broadly warp the composite S0 and S1 foliation in the quartzite layers, and tighter crenulation folds (148! N398W) in the mechanically weaker pelite layers (Fig. 16). The axial planes and fold axes of both the open F2 folds in the quartzite, and the crenulations in the pelitic layers, are parallel to each other (N508W, 648NE) indicating that they are part of the same deformation phase. Broad NE- and NW-trending subhorizontal folds in the Khumbu region have been documented by many previous studies (Hubbard 1988; Carosi et al. 1999a, b; Catlos et al. 2002; Searle et al. 2003) and are here termed the Khumbu Dome Complex (Fig. 3). An undeformed layer-parallel leucogranite sill, which is partially exposed above this outcrop on the Nuptse Lhotse wall, suggests that at least D1 and D2 predated its emplacement. Other leucogranite sills in the upper section of the GHS core also contain little evidence for solid-state fabric development, suggesting that much of the anatectic melting and leucogranite injection post-dated the polyphase folding that characterizes the core of the slab. Fig. 16. Photograph of key outcrop exposure of overprinting folds/fabrics used to define at least two phases of deformation in the core of the Greater Himalayan Slab. Uniform light grey layer is quartzite with bedding defined by biotite. Surrounding material is pelitic schist. See text for details. Brunton compass for scale. Image is of a vertical wall viewed towards the west. Approximate location of outcrop shown on Figure 3.